In context: Despite being highly effective at simulating meaningful exchanges between intelligent beings, large language models remain essentially deceptive software products with significant memory limitations. A startup is now attempting to sell a solution to the chatbots' memory problem and is willing to pay nearly a thousand dollars to a human reporter with a notoriously difficult relationship with technology.
As Boston Dynamics demonstrated years ago, "bullying" technology designed to mimic intelligent behaviors is nothing new. Memvid is now offering $800 to someone interested in putting modern AI models to the test – a "professional" yeller tasked with spending an entire day stressing popular chatbots.
The job posting seeks a single individual and is clearly designed to promote a solution to one of the most persistent issues in today's AI technology. Memvid, offering $100 per hour for the one-day position, emphasized that "AI memory" remains fundamentally broken.
Chatbots can briefly mimic intelligence, but they often "forget" what the user asked moments earlier. AI models are inherently prone to losing context across conversations, forcing users to repeat questions or restate facts repeatedly.
"That is the frustration behind this job. We are hiring someone to expose the problem in public by stress-testing chatbot memory all day long," Memvid's post notes.
The "Professional AI Bully" will be required to interact with popular AI chatbots, asking them to remember information and documenting the resulting chaos. The role demands detailed records of failed conversations, even though Memvid will also record the screen or the user via a camera.
The fully remote position does not require prior experience in AI "bullying," though candidates must provide an extensive personal history of negative experiences with technology. A strong opinion on the shortcomings of today's AI is also required.
According to Memvid co-founder and CEO Mohamed Omar, chatbots face a massive memory problem because early AI memory solutions were unreliable. The ability to retain information and maintain coherent conversations is the holy grail of a properly functioning chatbot, while hallucinations result from LLMs with poor memory "habits."
Memvid is marketing a self-improving memory layer for AI models: a single, file-based solution designed to make context persistent across conversations and to improve chatbot responses. The company claims it can enhance latency, accuracy, and infrastructure efficiency in AI products, either through its integrated chatbot solution, Kora, or via dedicated developer tools.
The AI memory problem is increasingly visible in real-world applications, with LLMs and chatbots showing growing reliability issues when handling massive data repositories. Memvid's job posting can be seen both as a clever advertisement for the company's products and as a reminder of the current limitations of AI. Omar confirmed that the search for the right "AI Bully" will conclude within a couple of weeks, with the company seeking someone who truly dislikes AI for the role.
